Pull Ceph updates from Sage Weil:
"There is a lot of refactoring and hardening of the libceph and rbd
code here from Ilya that fix various smaller bugs, and a few more
important fixes with clone overlap. The main fix is a critical change
to the request_fn handling to not sleep that was exposed by the recent
mutex changes (which will also go to the 3.16 stable series).
Yan Zheng has several fixes in here for CephFS fixing ACL handling,
time stamps, and request resends when the MDS restarts.
Finally, there are a few cleanups from Himangi Saraogi based on
Coccinelle"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client: (39 commits)
libceph: set last_piece in ceph_msg_data_pages_cursor_init() correctly
rbd: remove extra newlines from rbd_warn() messages
rbd: allocate img_request with GFP_NOIO instead GFP_ATOMIC
rbd: rework rbd_request_fn()
ceph: fix kick_requests()
ceph: fix append mode write
ceph: fix sizeof(struct tYpO *) typo
ceph: remove redundant memset(0)
rbd: take snap_id into account when reading in parent info
rbd: do not read in parent info before snap context
rbd: update mapping size only on refresh
rbd: harden rbd_dev_refresh() and callers a bit
rbd: split rbd_dev_spec_update() into two functions
rbd: remove unnecessary asserts in rbd_dev_image_probe()
rbd: introduce rbd_dev_header_info()
rbd: show the entire chain of parent images
ceph: replace comma with a semicolon
rbd: use rbd_segment_name_free() instead of kfree()
ceph: check zero length in ceph_sync_read()
ceph: reset r_resend_mds after receiving -ESTALE
...
Now that rbd_img_request_create() is called from work functions, no
need to use GFP_ATOMIC.
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Reviewed-by: Alex Elder <elder@linaro.org>
While it was never a good idea to sleep in request_fn(), commit
34c6bc2c91 ("locking/mutexes: Add extra reschedule point") made it
a *bad* idea. mutex_lock() since 3.15 may reschedule *before* putting
task on the mutex wait queue, which for tasks in !TASK_RUNNING state
means block forever. request_fn() may be called with !TASK_RUNNING on
the way to schedule() in io_schedule().
Offload request handling to a workqueue, one per rbd device, to avoid
calling blocking primitives from rbd_request_fn().
Fixes: http://tracker.ceph.com/issues/8818
Cc: stable@vger.kernel.org # 3.16, needs backporting for 3.15
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Tested-by: Eric Eastman <eric0e@aol.com>
Tested-by: Greg Wilson <greg.wilson@keepertech.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Currently, we use a rwlock tb_lock to protect concurrent access to the
whole zram meta table. However, according to the actual access model,
there is only a small chance for upper user to access the same
table[index], so the current lock granularity is too big.
The idea of optimization is to change the lock granularity from whole
meta table to per table entry (table -> table[index]), so that we can
protect concurrent access to the same table[index], meanwhile allow the
maximum concurrency.
With this in mind, several kinds of locks which could be used as a
per-entry lock were tested and compared:
Test environment:
x86-64 Intel Core2 Q8400, system memory 4GB, Ubuntu 12.04,
kernel v3.15.0-rc3 as base, zram with 4 max_comp_streams LZO.
iozone test:
iozone -t 4 -R -r 16K -s 200M -I +Z
(1GB zram with ext4 filesystem, take the average of 10 tests, KB/s)
Test base CAS spinlock rwlock bit_spinlock
-------------------------------------------------------------------
Initial write 1381094 1425435 1422860 1423075 1421521
Rewrite 1529479 1641199 1668762 1672855 1654910
Read 8468009 11324979 11305569 11117273 10997202
Re-read 8467476 11260914 11248059 11145336 10906486
Reverse Read 6821393 8106334 8282174 8279195 8109186
Stride read 7191093 8994306 9153982 8961224 9004434
Random read 7156353 8957932 9167098 8980465 8940476
Mixed workload 4172747 5680814 5927825 5489578 5972253
Random write 1483044 1605588 1594329 1600453 1596010
Pwrite 1276644 1303108 1311612 1314228 1300960
Pread 4324337 4632869 4618386 4457870 4500166
To enhance the possibility of access the same table[index] concurrently,
set zram a small disksize(10MB) and let threads run with large loop
count.
fio test:
fio --bs=32k --randrepeat=1 --randseed=100 --refill_buffers
--scramble_buffers=1 --direct=1 --loops=3000 --numjobs=4
--filename=/dev/zram0 --name=seq-write --rw=write --stonewall
--name=seq-read --rw=read --stonewall --name=seq-readwrite
--rw=rw --stonewall --name=rand-readwrite --rw=randrw --stonewall
(10MB zram raw block device, take the average of 10 tests, KB/s)
Test base CAS spinlock rwlock bit_spinlock
-------------------------------------------------------------
seq-write 933789 999357 1003298 995961 1001958
seq-read 5634130 6577930 6380861 6243912 6230006
seq-rw 1405687 1638117 1640256 1633903 1634459
rand-rw 1386119 1614664 1617211 1609267 1612471
All the optimization methods show a higher performance than the base,
however, it is hard to say which method is the most appropriate.
On the other hand, zram is mostly used on small embedded system, so we
don't want to increase any memory footprint.
This patch pick the bit_spinlock method, pack object size and page_flag
into an unsigned long table.value, so as to not increase any memory
overhead on both 32-bit and 64-bit system.
On the third hand, even though different kinds of locks have different
performances, we can ignore this difference, because: if zram is used as
zram swapfile, the swap subsystem can prevent concurrent access to the
same swapslot; if zram is used as zram-blk for set up filesystem on it,
the upper filesystem and the page cache also prevent concurrent access
of the same block mostly. So we can ignore the different performances
among locks.
Acked-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Reviewed-by: Davidlohr Bueso <davidlohr@hp.com>
Signed-off-by: Weijie Yang <weijie.yang@samsung.com>
Signed-off-by: Minchan Kim <minchan@kernel.org>
Cc: Jerome Marchand <jmarchan@redhat.com>
Cc: Nitin Gupta <ngupta@vflare.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
If we are mapping a snapshot, we must read in the parent_overlap value
of that snapshot instead of that of the base image. Not doing so may
in particular result in us returning zeros instead of user data:
# cat overlap-snap.sh
#!/bin/bash
rbd create --size 10 --image-format 2 foo
FOO_DEV=$(rbd map foo)
dd if=/dev/urandom of=$FOO_DEV bs=1M &>/dev/null
echo "Base image"
dd if=$FOO_DEV bs=1 count=16 skip=$(((4 << 20) - 8)) 2>/dev/null | xxd
rbd snap create foo@snap
rbd snap protect foo@snap
rbd clone foo@snap bar
rbd snap create bar@snap
BAR_DEV=$(rbd map bar@snap)
echo "Snapshot"
dd if=$BAR_DEV bs=1 count=16 skip=$(((4 << 20) - 8)) 2>/dev/null | xxd
rbd resize --allow-shrink --size 4 bar
echo "Snapshot after base image resize"
dd if=$BAR_DEV bs=1 count=16 skip=$(((4 << 20) - 8)) 2>/dev/null | xxd
# ./overlap-snap.sh
Base image
0000000: e781 e33b d34b 2225 6034 2845 a2e3 36ed ...;.K"%`4(E..6.
Snapshot
0000000: e781 e33b d34b 2225 6034 2845 a2e3 36ed ...;.K"%`4(E..6.
Resizing image: 100% complete...done.
Snapshot after base image resize
0000000: e781 e33b d34b 2225 0000 0000 0000 0000 ...;.K"%........
Even though bar@snap is taken with the old bar parent_overlap (8M),
reads from bar@snap beyond the new bar parent_overlap (4M) return
zeroes. Fix it.
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Currently rbd_dev_v2_header_info() reads in parent info before the snap
context is read in. This is wrong, because we may need to look at the
the parent_overlap value of the snapshot instead of that of the base
image, for example when mapping a snapshot - see next commit. (When
mapping a snapshot, all we got is its name and we need the snap context
to translate that name into an id to know which parent info to look
for.)
The approach taken here is to make sure rbd_dev_v2_parent_info() is
called after the snap context has been read in. The other approach
would be to add a parent_overlap field to struct rbd_mapping and
maintain it the same way rbd_mapping::size is maintained. The reason
I chose the first approach is that the value of keeping around both
base image values and the actual mapping values is unclear to me.
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Reviewed-by: Alex Elder <elder@linaro.org>
There is no sense in trying to update the mapping size before it's even
been set.
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Recently discovered watch/notify problems showed that we really can't
ignore errors in anything refresh related. Alas, currently there is
not much we can do in response to those errors, except print warnings.
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Reviewed-by: Alex Elder <elder@linaro.org>
rbd_dev_spec_update() has two modes of operation, with nothing in
common between them. Split it into two functions, one for each mode
and make our expectations more clear.
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Reviewed-by: Alex Elder <elder@linaro.org>
spec->image_id assert doesn't buy us much and image_format is asserted
in rbd_dev_header_name() and rbd_dev_header_info() anyway.
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Reviewed-by: Alex Elder <elder@linaro.org>
A wrapper around rbd_dev_v{1,2}_header_info() to reduce duplication.
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Make /sys/bus/rbd/devices/<id>/parent show the entire chain of parent
images. While at it, kernel sprintf() doesn't return negative values,
casting to unsigned long long is no longer necessary and there is no
good reason to split into multiple sprintf() calls.
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Free memory allocated using kmem_cache_zalloc using kmem_cache_free
rather than kfree. The helper rbd_segment_name_free does the job here.
Its position is shifted above the calling function.
The Coccinelle semantic patch that detects this change is as follows:
// <smpl>
@@
expression x,E,c;
@@
x = \(kmem_cache_alloc\|kmem_cache_zalloc\|kmem_cache_alloc_node\)(c,...)
... when != x = E
when != &x
?-kfree(x)
+kmem_cache_free(c,x)
// </smpl>
Signed-off-by: Himangi Saraogi <himangi774@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Since linux kernel 3.13, kthread_run() internally uses
wait_for_completion_killable(). We sometimes may use kthread_run()
while we still have a signal pending, which we used to kick our threads
out of potentially blocking network functions, causing kthread_run() to
mistake that as a new fatal signal and fail.
Fix: flush_signals() before kthread_run().
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
image_id is leaked if the parent happens to have been recorded already.
Fix it.
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Switch rbd_dev_header_{un,}watch_sync() to use the new helper and fix
rbd_dev_header_unwatch_sync() to destroy watch_request structures
before queuing watch-remove message while at it. This mistake slipped
into commit b30a01f2a3 ("rbd: fix osd_request memory leak in
__rbd_dev_header_watch_sync()") and could lead to "image still in use"
errors on image removal.
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Reviewed-by: Alex Elder <elder@linaro.org>
In the past, rbd_dev_header_watch_sync() used to handle both watch and
unwatch requests and was entangled and leaky. Commit b30a01f2a3
("rbd: fix osd_request memory leak in __rbd_dev_header_watch_sync()")
split it into two separate functions. This commit cleanly abstracts
the common bits, relying on the fixed rbd_obj_request_wait().
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Reviewed-by: Alex Elder <elder@linaro.org>
rbd_obj_request_wait() should cancel the underlying OSD request if
interrupted. Otherwise libceph will hold onto it indefinitely, causing
assert failures or leaking the original object request.
This also adds an rbd wrapper around ceph_osdc_cancel_request() to
match rbd_obj_request_submit() and rbd_obj_request_wait().
Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Alexander reported mkswap on /dev/zram0 is failed if other process is
opening the block device file.
Step is as follows,
0. Reset the unused zram device.
1. Use a program that opens /dev/zram0 with O_RDWR and sleeps
until killed.
2. While that program sleeps, echo the correct value to
/sys/block/zram0/disksize.
3. Verify (e.g. in /proc/partitions) that the disk size is applied
correctly. It is.
4. While that program still sleeps, attempt to mkswap /dev/zram0.
This fails: mkswap: error: swap area needs to be at least 40 KiB
When I investigated, the size get by ioctl(fd, BLKGETSIZE64, xxx) on
mkswap to get a size of blockdev was zero although zram0 has right size by
2.
The reason is zram didn't revalidate disk after changing capacity so that
size of blockdev's inode is not uptodate until all of file is close.
This patch should fix the BUG.
Signed-off-by: Minchan Kim <minchan@kernel.org>
Reported-by: Alexander E. Patrakov <patrakov@gmail.com>
Tested-by: Alexander E. Patrakov <patrakov@gmail.com>
Reviewed-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Cc: Nitin Gupta <ngupta@vflare.org>
Acked-by: Jerome Marchand <jmarchan@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Linus Torvalds
Joe Perches
Ilya Dryomov
Weijie Yang
Minchan Kim
Sergey Senozhatsky
Himangi Saraogi
Lars Ellenberg